Latino Representation in State Houses and Congress

Citation preview

Latino Representation in State Houses and Congress This book argues that Latino representation in U.S. legislative institutions is shaped not only by demographics, but also by legislative institutional design, as well as elite-driven methods, features of the electoral system, and the increasing mainstreaming of Latinos in American society. The election of Latino legislators in the United States is thus complex and varied. This book provides evidence on how successful Latinos have been in winning state legislative and congressional districts in which they have no natural advantage. In particular, this book demonstrates that Latino candidates benefit from higher percentages of Latino citizens in the state, more liberal citizenries, and citizen legislatures. Jason P. Casellas argues that the legislatures most conducive to the election of Latino candidates are Florida, New Mexico, and California, whereas the least conducive are the U.S. House and New York. Jason P. Casellas is Assistant Professor of Government and Associate Director of the Irma Rangel Public Policy Institute at the University of Texas at Austin. He specializes in American politics, with specific research and teaching interests in Latino politics, legislative politics, and state and local politics. He is the recipient of numerous fellowships, including a Princeton President’s Fellowship, an American Political Science Association Fellowship, and a Ford Motor Company Fellowship. His dissertation won third place in a nationwide interdisciplinary competition for the best dissertation given by the American Association of Hispanics in Higher Education and the Educational Testing Service. In 2007–2008, he was the Samuel DuBois Cook Postdoctoral Fellow at Duke University. In 2009–2010, he was a visiting postdoctoral Fellow at the United States Studies Centre in Sydney, Australia. His work has appeared in the Journal of Politics, Legislative Studies Quarterly, Political Research Quarterly, Qualitative Methods, and the Journal of California Politics and Policy.

Latino Representation in State Houses and Congress

JASON P. CASELLAS University of Texas, Austin

cambridge university press Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo, Delhi, Dubai, Tokyo, Mexico City Cambridge University Press 32 Avenue of the Americas, New York, ny 10013-2473, usa www.cambridge.org Information on this title: www.cambridge.org/9780521198974 © Jason P. Casellas 2011 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2011 Printed in the United States of America A catalog record for this publication is available from the British Library. Library of Congress Cataloging in Publication data Casellas, Jason Paul, 1977– Latino representation in state houses and Congress / Jason P. Casellas. p. cm. Includes bibliographical references and index. isbn 978-0-521-19897-4 (hardback) 1. Legislative bodies – United States. 2. Legislative bodies – United States – States. 3. United States. Congress. 4. Hispanic American legislators. 5. Hispanic Americans – Politics and government. I. Title. jk2495.c37 2011 328.730089′68–dc22 2010031213 isbn 978-0-521-19897-4 Hardback Cambridge University Press has no responsibility for the persistence or accuracy of urls for external or third-party Internet Web sites referred to in this publication and does not guarantee that any content on such Web sites is, or will remain, accurate or appropriate.

To my parents

Contents

List of Tables List of Figures Acknowledgments Introduction Latinos in American Society Latino Political Incorporation The Concept of Representation Brief Discussion of Subsequent Chapters

1

Latinos in Legislatures: Historical and Theoretical Setting Literature on Latinos in Legislatures Other Minorities in Legislatures and Redistricting Research Questions Types of Data Choice of Legislatures

2

The Effects of Population, Turnover, and Term Limits on Latino Representation Turnover, Professionalization, and Term Limits Methods Institutional and Demographic Determinants of Latino Representation Conclusion

3

District Composition and the Election of Latino Candidates Redistricting after the Voting Rights Act of 1965 Methods United States House

page x xiii xv 1 4 9 10 15 19 22 25 28 30 32 33 34 41 43 46 51 52 56 57 vii

Contents

viii Overall Findings New Mexico California Texas Arizona Florida New York New Jersey Conclusion

4

Electing Latinos in Non-Latino Majority Districts Methodology The Elite-Driven Process Features of the Electoral System Latino Republicans Conclusion

5

Voices from Within: How Latino Legislators See Themselves Methodology Background, Political Past, and Election to Current Position Legislators’ Perceptions of Competitiveness How They View Their Districts Issue Priorities of Latino Legislators Perceptions of Representation Does Partisanship Trump Ethnicity? Conclusion

6

Roll Call Voting Behavior of Latino Legislators Background of Latinos in Colorado, New Jersey, and Texas What We Know About Latino Roll Call Voting Behavior Data and Methods Findings Conclusion

7

Conclusion Legislatures and Legislators Matter Not All States Are Equal: Institutions and Demographics Matter The Mainstreaming of Latinos in U.S. Legislatures African Americans and Latinos Party Outreach Extends Beyond Presidential Races Latino Pan-Ethnic Identity Despite District Differences Variations in Latino Legislators’ Voting Records

59 63 64 65 67 68 70 72 73 76 78 82 91 101 103 106 108 110 112 113 115 118 120 122 125 126 128 130 131 136 138 139 140 142 143 143 145 145

Contents Appendix A Appendix B Appendix C Appendix D Appendix E References Index

ix

147 149 157 159 161 163 175

Tables

1.1a 1.1b

1.2 1.3 1.4 2.1 2.2 3.1 3.2 3.3 4.1a 4.1b 4.2

x

Latino Representatives in the U.S. Congress Ranked in Descending Order of Latino Population, 2009 Non-Latino Representatives in the U.S. Congress Ranked in Descending Order of Latino Population (40 Percent and Above Districts), 2009 Percentage Latino in Seven States and United States, 1950–2000 (Census Bureau) Percentage Latino in Seven State Legislatures and Congress, 1986–2002 Literature on Racial and Ethnic Representation in U.S. State Legislatures Term Limits Legislation in the United States The Effects of Turnover and Term Limits on Latino Representation, 1992, 1996, 2000, and 2004 Probit Analysis of Latino Representation in State Legislatures and the U.S. House Probability of a Latino Being Elected to Eight Legislatures, 2003–2004 Citizens Probability of a Latino Being Elected to Eight Legislatures, 2003–2004 All Latino Representatives in the United States, 2004 African-American Representatives in the United States, 2004 Two Minority Candidates in General Election

page 20

21 21 22 23 36 44 58 60 61 79 80 85

Tables 4.3 4.4 4.5 4.6 4.7 4.8 5.1 6.1

A.1

xi

Latino Legislators First Appointed to Seat Latino Legislators Recruited by Party Leaders Latino Legislators with Anglo Names Latino Legislators in Near-Majority Districts (45 Percent and Up) Latino Legislators in Districts with Combined Latino and African-American Majorities Latino Legislators Who Initially Won in Multicandidate Primaries Descriptive Statistics of Interview Subjects Impact of Latino Representative and Percent Latino on Poole-Rosenthal Scores for Texas House, Colorado Legislature, New Jersey Legislature, and U.S. House, 1999–2002 Probit Analysis of Latino Representation in New Mexico’s Lower Chamber and Entire Legislature, 1991 and 2003

Probability of a Latino Being Elected to the 2003 New Mexico Legislature A.3 Probit Analysis of Latino Representation in California’s Lower Chamber and Entire Legislature, 2000 and 2003 A.4 Probability of a Latino Being Elected to the 2003 California Legislature A.5 Probit Analysis of Latino Representation in Texas’s Lower Chamber and Entire Legislature, 2001 and 2003 A.6 Probability of a Latino Being Elected to the 2003 Texas Legislature A.7 Probit Analysis of Latino Representation in Arizona’s Lower Chamber and Entire Legislature, 2001 and 2003 A.8 Probability of a Latino Being Elected to the 2003 Arizona Legislature A.9 Probit Analysis of Latino Representation in Florida’s Lower Chamber and Entire Legislature, 1996 and 2003 A.10 Probability of a Latino Being Elected to the 2003 Florida Legislature A.11 Probit Analysis of Latino Representation in New York’s Lower Chamber and Entire Legislature, 2003

87 90 92 94 95 99 124

135

149

A.2

149

150 150 151 151 151 152 152 152 153

xii

A.12 Probability of a Latino Being Elected to the 2003 New York Legislature A.13 Probit Analysis of Latino Representation in New Jersey’s Lower Chamber and Entire Legislature, 2003 A.14 Probability of a Latino Being Elected to the 2003 New Jersey Legislature A.15 Probit Analysis of Latino Representation in the U.S. House, 1995 and 2003 A.16 Probability of a Latino Being Elected to the 2003 U.S. House A.17 Probit Analysis of Latino Representation in Seven State Lower Chambers and the U.S. House, 2003–2004 A.18 Probit Analysis of Latino Representation in Five State Lower Chambers and the U.S. House, Pre-2000 Redistricting A.19 Probit Analysis of Latino Representation in Seven State Legislatures, 2003–2004 A.20 Probit Analysis of Latino Representation in Five State Legislatures, Pre-2000 Redistricting A.21 Impact of Mexican-American and Puerto Rican Representative and Percent Latino on Poole-Rosenthal Score for the 100th Congress A.22 Impact of Cuban-American, Mexican-American, and Puerto Rican Representative and Percent Latino on Poole-Rosenthal Score for the 87th–104th Congresses

Tables

153 153 154 154 154 155

156 156 159

161

162

Figures

2.1 2.2 2.3 2.4 2.5 2.6

Latino representation in state lower houses, 1992 Latino representation in state upper houses, 1992 Latino representation in state lower houses, 2000 Latino representation in state upper houses, 2000 Latino representation in state lower houses, 2004 Latino representation in state upper houses, 2004

page 35 35 48 48 49 49

xiii

Acknowledgments

Words cannot express my appreciation to the many people who have given me the encouragement, advice, and invaluable criticisms that have shaped this book. First and foremost, I am most indebted to my dissertation advisor, R. Douglas Arnold, for always being supportive of my work and helping me become a better scholar and writer. His numerous and always detailed critiques of my drafts have undoubtedly made this book much better than it otherwise would have been. I have always been amazed by Doug’s ability to provide timely and, more importantly, quality reactions to my work. This never ceased during my time at Princeton despite a challenge to Doug’s health. Secondly, Tali Mendelberg has been a strong supporter of my project ever since we first talked about it in her office early on during my graduate career. Tali has had faith in my abilities since she first read my application file in 2000, and she has continued her support of my work and provided important feedback despite her taking on important departmental responsibilities as Director of Graduate Studies. Her detailed and incisive commentaries on my work no doubt have made this book much clearer and systematic. I am forever indebted to Tali for her valuable mentorship. Larry Bartels has also served as an important advisor, especially in regards to questions I have had regarding the quantitative aspects of my book. Larry’s knowledge of redistricting, especially in New Jersey, helped inform my book. I am also indebted to other members of the Princeton faculty, including those in attendance at the American Politics Research Seminar where I presented drafts of this project. The staff at CSDP, including Helene Wood and Diane Price, provided years of support, laughter, and surrogate motherly care during my years in graduate school. Monica Selinger at Princeton made my life xv

xvi

Acknowledgments

a breeze by always looking out for me and enduring my long conversations in her office about topics ranging from where to go in Italy to the ins and outs of navigating the Princeton bureaucracy. I also thank Princeton University’s Graduate School, especially Dean David Redman and Danielle Gray, for having confidence in me and asking me to represent the school at various recruiting fairs across the country as a Student Diversity Assistant. I would also like to thank the Horowitz Foundation for Social Policy for a dissertation research grant, which made some of the research presented possible, as well as the Earhart Foundation and Robby George for supplemental financial support during my years at Princeton. Finally, I would also thank John Darley and the Fellowship of Woodrow Wilson Scholars at Princeton University that provided much needed financial and intellectual support at the most critical time of my dissertation writing. Although I organized and collected most of the data used in this project, I thank NALEO, the Secretaries of State for the seven states of my study, David Lublin, and Gary Moncrief for providing important data. I especially thank Nolan McCarty for generously providing state NOMINATE data for the Colorado, New Jersey, and Texas legislatures. I also thank my colleagues at the University of Texas at Austin, for their support and confidence in my abilities, especially John Higley and Gary Freeman, who as chairs of the department have supported my research in many ways. I also thank the University of Texas at Austin Graduate Studies program for generously providing a Summer Research Assignment in 2007, which assisted me in completing a great deal of the research in this book. David Leal has also served as an invaluable mentor to me as a young assistant professor. His leadership of the Public Policy Institute has made my job as the Associate Director flawless and, most important of all, painless. Daron Shaw, Bat Sparrow, and, more recently, Nick Valentino have been wonderful colleagues at UT. I also thank my other faculty colleagues at UT, including Brian Roberts, Tasha Philpot, Eric McDaniel, Ismail White, Corrine McConnaughy, Pat McDonald, Jason Brownlee, Ken Greene, Sean Theriault, and Andy Karch. I especially thank Paula McClain, Kerry Haynie, Mike Munger, John Aldrich, Alisa Kessel, Efren Perez, Niambi Carter, Dick Engstrom, and Mary Bogues for making my time at Duke University’s Social Science Research Institute as the first Samuel DuBois Cook Postdoctoral Fellow a valuable and rewarding experience. In particular, I thank Rodney Hero and David Lublin for participating in a Book Manuscript Conference at UT, where they provided excellent feedback on this project. I am also grateful to the U.S. Studies Centre at the University of Sydney, including

Acknowledgments

xvii

Margaret Levi,Geoff Garrett, and Brendon O’Connor, where I completed some of the last-minute edits to this book. I also thank the many anonymous reviewers for providing valuable feedback. Most of all, I thank my family and friends and their prayers for helping me through graduate school and my first few years as an assistant professor. Without them, I could not have survived.

Introduction

In recent political campaigns, candidates have adopted Spanish-language appeals in their efforts to woo the growing Latino vote. The reasons for such appeals are obvious: Latinos are the largest minority group in the United States, and the Pew Hispanic Center estimates that the Latino population will increase from 14 percent of the population in 2007 to nearly 30 percent of the population in 2050.1 However, although the growth of Latinos in American legislative institutions has slowly increased in the past decade, Latinos remain underrepresented in Congress and state legislatures. Even though Congress has long been the focal point for studies of representation, a comparative analysis of Congress and state legislatures has yet to be done. This book is the first systematic examination of the election of Latinos to U.S. state legislatures and Congress. This book argues that Latino representation is dependent on subethnic diversity, distinct political backgrounds even among Mexican Americans, and nascent political experience. The central argument is that Latino representation in U.S. legislative institutions is shaped not only by demographics, but also by legislative institutional design, as well as elite-driven methods, features of the electoral system, and the increasing mainstreaming of Latinos in American society. The election of Latino legislators in the United States is thus complex and varied. Through specifying the political processes and mechanisms of Latino political incorporation, the central questions this book addresses are as follows: How do we explain the election of Latino candidates to Congress 1

See the Pew Hispanic Center website for reports on the growing Latino population in the United States (http://www.pewhispanic.org).

1

2

Latino Representation in State Houses

and state legislatures? Do Latino candidates have unique obstacles being elected to legislative office? To what extent have Latinos benefited from the creation of majority-Latino districts in order to be elected? Are Latinos a special group or just another of the many ethnic groups in the American melting pot? Are Latino candidates as advantaged when they run in districts filled with citizens who share their ethnic heritage as other ethnic candidates have been throughout American history? Are they seriously disadvantaged when they run in heterogeneous districts? How do Latino representatives see themselves? What difference does it make whether Latino legislators represent Latinos? Do Latino representatives behave differently than non-Latino representatives? This book provides answers to all of these questions. Why are these questions important in the first place? Especially since the Voting Rights Act of 1965, scholars, pundits, and political observers have debated issues regarding race, redistricting, and legislative representation. For example, the Supreme Court weighed in on the Voting Rights Act in a 2009 ruling that overturned a lower court’s decision to require an Austin, Texas utility district to comply with Section 5 of the Act, which requires the Department of Justice to “preclear” any changes in voting procedures. Chief Justice John Roberts wrote that Section 5 “raises serious constitutional concerns” that can only be justified by “current needs” rather than the legacy of racial discrimination.2 This book provides evidence to show how successful Latinos have been in winning state legislative and congressional districts in which they have no natural advantage. In particular, across seven diverse states and Congress, this book demonstrates the institutional and demographic determinants of Latino representation, as well as the extent to which Latino legislators see themselves as distinctive representatives. These questions are also especially important as the number of Latinos in American society continues to grow every day. Latinos currently constitute 14 percent of American society, and this percentage is estimated to reach 18 percent by 2025.3 Should we expect that the flow of Latino candidates into Congress and state legislatures will increase at comparable levels as Latino citizens increase in numbers and put down their

2 3

Qtd. In The Economist, June 27, 2009. The number of Latinos in the United States reached 44.3 million on July 1, 2006, accounting for about one-half of people added to the nation’s population since July 1, 2005. These estimates do not include the island of Puerto Rico (from U.S. Census American Fact Finder).

Introduction

3

political roots? Or to what extent does the election of Latino candidates rest on governments drawing legislative districts that favor their election? Although these questions are important, they are not the ones scholars have investigated thoroughly. The second reason that questions about Latino representation are important is that they reflect underlying questions about race and ethnicity broadly conceived. Some people assume that the story of Latino candidates parallels the story of African-American candidates. According to this view, the only way to increase the number of Latino legislators is to create districts with Latino majorities. Although majority-African-American districts were apparently necessary for the election of substantial numbers of African-American legislators, there are reasons to believe that the story might not be quite the same for Latino candidates. First, many Latinos have not suffered centuries of discrimination.4 Whatever discrimination Latino citizens suffer is probably somewhere between that of other volunteer immigrants – for example, Irish, Italian, and Polish Americans – and the discrimination imposed on people brought to the country in chains (Erie 1988; Skerry 1993). Second, Latino citizens have more choices than do African-American citizens about whether to identify strongly with their heritage or to hide it by diving into the melting pot. Ethnic self-identity is often a choice; skin color is not. As Dawson (1994) and Pinderhughes (1987) show, African Americans have experienced difficulties regarding political incorporation precisely because of this history of American racism based on skin color. Likewise, Massey and Denton (1993) make clear that individuals with black skin color face more discrimination than Latinos and Asians. At the same time, there is no question that Latinos have experienced discrimination in many ways (Hero 1992; Montejano 1999). This book goes beyond previous studies of Latino representation, which only focus on Congress. Such an analysis offers significant advantages. First, it expands the number of Latino politicians to be studied. Because only twenty-eight Latinos currently serve in Congress, focusing exclusively on how those politicians gained office does not provide much leverage on the central questions. Approximately 220 Latinos serve in the fifty state legislatures, which provides a larger pool of subjects. Second, broadening the study to include state legislators increases the variance 4

This does not suggest that Latinos have not suffered discrimination. In particular, stories abound about discrimination in the Southwest and even South Florida where real estate signs compared Mexicans and Cubans to dogs.

4

Latino Representation in State Houses

in the explanatory variables. Some states have high concentrations of Latinos; most do not. Some states have been more aggressive than have others in creating majority-Latino districts. Some states elect two or more legislators from the same districts. This increased variance makes it easier to assess the impact of demographic patterns, districting arrangements, and electoral laws on the election of Latino candidates. Third, serving as a state legislator often provides a stepping-stone toward congressional election. Women were not elected to Congress in large numbers until they first succeeded in state legislatures. Examining Latino representation in state legislatures offers a window on the future of Latinos in Congress. Finally, state legislatures are themselves important. One cannot understand domestic policy making without appreciating the role of state governments and state legislatures. Latinos in American Society Latinos are the fastest-growing minority group in the United States and have surpassed African Americans as the largest minority group in the country. Latinos are also a diverse group. They differ markedly in their geographic origins, their length of residence, and their identification with being Latino. Many Mexican Americans, for example, have resided in the Southwest for hundreds of years. Their unique cultural history is part of the political culture in states such as Texas, California, or New Mexico. Most Puerto Rican and Cuban immigrants, however, have more recently immigrated to the mainland United States for economic and political reasons, respectively. For many years, the Census merely collected data on those with “Spanish surnames,” but today, the Census collects data on country of origin. It was not until the 1960s that the number of Latino immigrants swelled, making the United States the fifth-largest Spanishspeaking country in the world. Unlike African Americans, Latinos are less reliably tied to the Democrats. Although only 20 percent of Latinos nationwide are registered Republican, President George W. Bush received approximately 35 percent of the Latino vote in the 2000 election. Foreign-born Latinos, second-generation Latinos, and third-generation Latinos classify themselves somewhat differently. The longer a Latina is in the United States, the more likely she is to be a Democrat. Foreign-born Latinos are the most likely to identify themselves as Independent, which helps explain why political campaigns have used Spanish-language advertising

Introduction

5

in many battleground states. Fourteen percent of third-generation Latinos call themselves Republicans, whereas a statistically indistinguishable 11 percent of foreign-born Latinos call themselves Republican. Since approximately 35 percent of Latinos are independent, both political parties have recently campaigned for their support. A closer look at the history of Mexican Americans, Puerto Rican Americans, and Cuban Americans illustrates the sociological and political diversity of Latinos. Mexican Americans Mexican Americans are the largest and oldest Latino group in the United States and comprise the majority of Latino legislators. More than twothirds of Latinos living in the United States are Mexican American. Much of the Southwest was part of Mexico until the Treaty of Guadalupe Hidalgo in 1848 (Gann and Duignan 1986: 17). This treaty ceded territory that eventually became eight states. Many Mexicans lived on this land and remained even after control switched from Mexico to the United States. Once non-Latino Americans settled in these territories, racial friction and prejudice emerged. Many Americans viewed Mexicans as “cowardly, ignorant, lazy, and addicted to gambling and alcohol” (Gann and Duignan 1986: 14). Mexican Americans currently represent one-quarter of the electorate in such crucial states as Texas and California. Mexican Americans have traditionally been loyal members of the Democratic Party, although in places like New Mexico, Republicans have long worked to incorporate Latinos into their fold. In New Mexico, approximately 16 percent of Latinos are registered Republicans. The number of “not so strong Democrats” nationwide is nearly 30 percent for Mexican Americans, more so than any other Latino group. Mexican Americans are more likely to vote for Democratic candidates than Republicans but are not as strong in their support of the Democratic Party as Cuban Americans are of the Republican Party.5 Not only are Mexican Americans the oldest and most established Latino group in the United States, but they are also the fastest-growing Latino group. Whereas most Mexican Americans originally settled in California and the Southwest, recent waves of Mexican immigrants have settled all across the United States, especially in the Midwest and South. In fact, the Tomás Rivera Policy Institute reports that nine of the top 5

See Latino National Survey (2008) for further details on Latino voter attitudes.

6

Latino Representation in State Houses

ten fastest-growing Latino counties in the United States are in the states of Arkansas, Georgia, and North Carolina.6 The growth of Mexican migrant workers is not limited to the South, either. Even industrial areas in the Midwest have experienced a rapid growth of Mexican workers in recent years. Puerto Rican Americans Puerto Ricans are unique among Latinos. They are United States citizens whether they reside in Puerto Rico or settle in any of the fifty states. Unlike other Latinos, they are never “illegal” immigrants. Puerto Ricans comprise about 9 percent of the United States Latino population, and most of them have traditionally settled in New York City. Puerto Ricans first migrated to New York following the Spanish American War. It was not until after World War II that Puerto Ricans migrated en masse to New York (Gann and Duignan 1986: 78). Puerto Rican immigration to the United States has been attributed to “the economic policies of the United States and the island’s government, which have encouraged industrialization and capitalist investment” (Hero 1992: 39). Like today’s Mexican immigrants, Puerto Ricans have come to the mainland United States for economic improvement and prosperity. More recently, however, Puerto Ricans have increasingly chosen central Florida as a preferred destination. According to the Puerto Rican Federal Affairs Administration, nearly 650,000 Puerto Ricans reside in Florida. In 2002, then-Governor Jeb Bush won 55 percent of the non-Cuban Latino vote (which includes a large portion of Puerto Ricans), according to Democratic pollster Sergio Bendixen. Despite the more partisan split in Florida among Puerto Ricans, most have nonetheless been strong supporters of the Democratic Party, and major Puerto Rican members of Congress have been strong Democrats. Approximately 50 percent of Puerto Ricans are registered Democrats.7 Puerto Ricans tend to support increased government support for the poor and see the Democratic Party as more willing to care and implement programs designed to help the economically disadvantaged. In recent elections, Puerto Ricans have supported Democratic candidates in large numbers. In 1996, former President Clinton captured 93 percent of the Puerto Rican vote. Nevertheless, some Republican candidates have found 6 7

See Tomás Rivera Policy Institute website for precise numbers (http://www.trpi.org). See Pew Hispanic Center/Kaiser Foundation National Survey of Latinos: The Latino Electorate, 2002 (http://www.pewhispanic.org/site/docs/pdf/latino_chartpack_092002.pdf).

Introduction

7

ways to appeal to Puerto Rican voters. Former Mayor Rudolph Giuliani (R-New York) captured 37 percent of the Puerto Rican vote in his successful mayoral re-election campaign. Mayor Michael R. Bloomberg (I-New York) won his 2001 race with strong support from New York City’s Latino community, which is mostly Puerto Rican and Dominican. Dominicans have consistently supported the mayor at higher levels than Puerto Ricans, though. Bloomberg was also perceptive in appealing to Dominicans, who are the fastest-growing immigrant group in parts of New York City (e.g., The Bronx). One could argue that these numbers merely indicate the failure of the Republicans at the national level and the success of Giuliani and Bloomberg at the local level. The better lesson is that Puerto Ricans are not “Yellow Dog” Democrats: They respond to candidate appeals, not only to party loyalty. Cuban Americans Most Cuban Americans came to the United States following Fidel Castro’s revolution of 1959. Cuban Americans comprise only 4 percent of the United States Latino population, and the largest number settled in the Miami area. Cuban Americans came to the United States more for political refuge than for economic opportunities. Gann and Duignan note that “when many Mexican or Puerto Rican intellectuals turned to antiestablishment politics in the United States, most Cubans looked upon the United States as a refuge against tyranny” (1986, 110). This distinction is crucial when determining why most Cuban Americans vote Republican and why most Cuban members of Congress and state legislators are also Republican. Traditionally, the Republican Party has been more supportive than the Democratic Party of tightening the trade embargo on Cuba. In addition, many Cuban Americans perceived the Republican Party as more anticommunist during the Cold War. More than two-thirds of Cuban Americans oppose U.S. relations with Cuba. Forty-eight percent of Cubans surveyed classified themselves as strong Republicans. This figure exceeds the number of Puerto Ricans and Mexican Americans who claim to be strong Democrats by approximately ten percentage points and sixteen percentage points, respectively.8 In addition, Cuban Americans born in Cuba during the Castro years are less likely to be Republican because they were not subjected to the discourse of “el exilio” in heavily Republican Miami. Cuban Americans who 8

See Pew Hispanic Center/Kaiser Foundation National Survey of Latinos: The Latino Electorate, 2002 (http://www.pewhispanic.org/site/docs/pdf/latino_chartpack_092002.pdf).

8

Latino Representation in State Houses

speak at least some Spanish are more likely to be Republican, especially those who were born in Cuba. This has to do with the large number of Cuban-born exiles who left Cuba in 1980 during the Mariel boatlift, whose political attitudes are not as hardline as those who left immediately following the revolution. It appears then that immigrants who left Cuba after 1959 but before the Mariel boatlift in 1980 are more likely to identify as partisan Republicans. In 2000, the Elián González controversy ignited the passions of many Cuban Americans and their congressional representatives who believed that the Clinton Administration’s Department of Justice, headed by Janet Reno, improperly sent the young boy back to an oppressive tyrannical regime. Many Cuban Americans who themselves left behind mothers, fathers, and siblings in Cuba in the early 1960s saw this incident as an affront to the principle that the ability to live in freedom easily outweighs the rights of a father. Al Gore broke with President Clinton and believed that the boy should have been allowed to stay in the United States. Despite this position, Gore received fewer Cuban American votes in Florida in 2000 than did Clinton in 1996, thus costing him Florida’s crucial electoral votes. In 2004, President Bush won the state of Florida in the general election by a comfortable margin. The presence of a Cuban American, Mel Martínez, in the U.S. Senate race had a reverse coattail effect by which Cuban American voters turned out in large numbers for the Republican ticket. Overall, Bush received 54 percent of the Latino vote, whereas Martínez received 59 percent of the Latino vote in Florida. Among Cuban Americans, Bush and Martínez enjoyed over 80 percent of the vote, according to CNN exit polls. In Florida, Cuban Americans are an influential voting bloc, even though other Latino groups comprise the majority of the Latino population.9 This brief history is by no means exhaustive, but when I refer to “subethnic” differences throughout the book, I am referring to differences among the three major Latino groups. Because Mexican Americans, Puerto Ricans, and Cuban Americans are often lumped together as “Latinos,” it behooves us to think carefully about context, region, skin color, and historical background within and across the group.

9

According to the U.S. Census, Cuban Americans account for 5 percent of the population in Florida. Latinos as a whole comprise 16 percent of the population. Other Latino groups comprise 6 percent of the entire state population, whereas Puerto Ricans comprise 3 percent and Mexican Americans 2 percent.

Introduction

9

Latino Political Incorporation What is the best theoretical way to understand Latino political incorporation? Moreover, at what point does a mainstream entrance into the political process replace ethnic-based politics? Early scholars such as Dahl (1961) argued that ethnic minorities initially are bound together by common characteristics and coalesce for political action, often by party elites eager to bring in more voters. In New Haven, the newer Italian American immigrants were courted by the Republican Party in opposition to the Irish-controlled Democratic Party. In a similar way, some Republicans have tried to court Latinos away from the Democratic Party, which has been the party most Latinos identify with. In the past several presidential elections, both political parties began appealing to Latino voters through Spanish-language media and targeted outreach efforts. As a result, more and more Latinos feel empowered to the point of having a stake in the political process (Rogers 2006). The questions then become: Are Latinos beginning to shed their ethnic loyalties and think of themselves as American? And do non-Latinos begin to see Latinos as American? As this transformation develops, we will presumably begin to see more Latinos elected to office from districts with non-Latino majorities. The process of assimilation often develops over the course of two generations, as many sociologists have observed (Alba and Nee 2003). To what extent will Latinos follow the path of Europe’s ethnic immigrants or the more difficult incorporation of African Americans? As Dawson (1994) has shown, the pluralist model does not account for racial discrimination and prejudice. African Americans still suffer disproportionately from poverty and other social ailments, just as many Latino groups do. Accordingly, Hero (1992) articulated a theory of “two-tiered pluralism” to explain the Latino experience in America, which acknowledges the history of discrimination and subordination but contends that pluralism best explains Latino political incorporation (although not as successful as for white ethnics). As we will see with many of the Latino legislators elected to non-Latino majority districts, a “Latino” identity is often not acknowledged or made known to voters. Many of these Latino representatives see themselves as regular Americans representing their districts. Before turning to a more in-depth discussion of Latino representation, it might be useful to review how political scientists have understood the concept.

10

Latino Representation in State Houses

The Concept of Representation Empirical scholars in political science have generally allowed normative theorists to conceptualize key concepts such as democracy, accountability, and representation (Collier and Adcock 1999; Pitkin 1967). For some empiricists, taking the time to revisit the very concepts they are purportedly measuring and testing seems at best too philosophical, and hence out of their domain, although Goertz (2006) offers a comprehensive treatment of social science concepts. Consequently, the empirical literature on representation has focused too heavily on statistical roll call analyses, which to a certain degree can help us ascertain the extent to which legislators represent their constituents in legislatures and Congress. Substantive representation, however, involves much more than how legislators vote. In order for political scientists to understand why, we must think carefully about what representation involves in terms of concepts, typologies, and case selection. Normative Conceptions of Representation Starting with Pitkin (1967), political scientists have regarded representation as either descriptive or substantive. Descriptive representation refers to citizens being represented by legislators who share particular demographic characteristics (race, gender, or ethnicity), whereas substantive representation involves legislators representing citizens’ interests or particular preferences. Scholars of black representation have debated the merits of which type of representation is most effective, with Swain (1993) arguing that substantive representation is what really counts whereas Mansbridge (1999) places more value on descriptive representation. Mansbridge, however, is a normative theorist who has argued that descriptive representation is essential for advancement of minorities and women in the American political system. Pitkin’s analysis did not really deal with minority representation, but the concept she presented has been extended to such studies. To date, no work on racial representation has challenged Pitkin’s conceptual framework or analyzed the concept of representation using more recent empirical research on methods.10

10

This is not the case regarding gender representation. See Celis (2008) for a thorough review of representation from a women’s studies perspective.

Introduction

11

The Concept(s) What does it mean to be represented? What does it mean to represent others? We are always asking others to represent us before a lawmaking body, a Court, or other institution. When one is represented by an attorney before a court, the attorney acts in the material interests of her client. It does not matter whether the attorney looks like her client physically. As long as the attorney defends her client well, then the client will be satisfied. In terms of political representation, however, surely more is at stake than just material interests. As Aristotle observed long ago, political issues deal with how we ought to order our lives together in the larger community. Additionally, politics deals with how individuals will be treated, including policies such as affirmative action, immigration, and English Only laws, just to name a few. This distinction is crucial because when Pitkin talks about descriptive and substantive representation, she is referring to political representation. As Gerring explains, concepts are not static and “progress in the cultural sciences occurs, if it occurs at all, through changing terms and definitions” (Gerring 2001). Issues of race and representation have dramatically changed since Pitkin’s exegesis, which requires conceptual revisions. All too often, social scientists have viewed variables as independent of each other. With regard to representation, scholars have dichotomized representation by splitting the concept into descriptive and substantive. Scholars such as Charles Ragin would argue for a configuration where these two concepts are placed in a spatial continuum in a way that acknowledges the diversity, albeit limited, of the concept of representation. As Ragin explains, membership in sets is “often partial” and rarely do we find cases that are either in or out of a given category (Ragin 1999). In the area of representation, most empirical researchers have chosen to use Pitkin’s dichotomy to test whether given groups are being adequately represented. For the most part, scholars in this research tradition have argued that substantive representation is what really matters, although many scholars argue that there is intrinsic value in descriptive representation based on issues of justice such that certain groups, especially women, should have some parity in political institutions (Phillips 1995; Young 1997). These matters are important in that many of the arguments regarding which type of representation matters deal more with conceptualization than with how representation per se is measured. Hardly anyone would disagree that descriptive representation without substantive representation

12

Latino Representation in State Houses

is not worth very much. However, the real question is whether those who represent districts with different demographic characteristics can adequately represent their constituents despite the physical difference. In this sense, the scholarship on racial representation suffers because of a lack of attention to the conceptualization of representation.11 Typologies How should representation be conceptualized? The previous section offered a critique of the prevailing norms of research regarding the issue of political representation. Given the complexity of this concept, it is fitting to explore the many different cells that comprise the “property space” of the concept of representation.12 A typology is a device for “partitioning events into types that share specified combinations of factors” (Stinchcombe 1968). Typologies can be complex or simple. Typologies for democracy can become quite complex, whereas typologies for approval ratings for the President are quite simple. In the area of representation, I am arguing that the prevailing typology is too simple, and that we need to complicate this concept further into a more encompassing typology, much in the same way Elman (2005) has done for the study of international politics. As discussed in the previous section, Ragin has offered a framework that assesses the importance of typologies and their implications. Ragin rightly points out that many researchers do not know where to proceed once their typology is fleshed out. In the tradition started by Lazarsfeld’s property space typology construction, Ragin sets forth his configurational approach as the heir to Lazarsfeld. Through functional reduction, it is argued, researchers can narrow the domain of researchable cells. In the case of racial representation, it is conceivable to imagine various scenarios of elected representatives and their demographic characteristics, such as black, urban, Democratic members from majority-Latino districts, and so on. For example, in Chapter 4, using Mill’s methods, I identify the commonalities and differences among Latino legislators elected to represent districts that are not majority-Latino. Gerring offers insight into precisely how to create a typology for social scientific concepts. In his analysis of ideology, Gerring notes 11

12

This is not the case in the literature on gender and representation, where there has been more of an effort to address these important issues. See Phillips (1998), Young (1997), and Celis (2008) for explicit treatments of gender and representation. See Lazarsfeld and Barton (1951) for more on the issue of property spaces in political science research.

Introduction

13

that typologies are usually created in several ways. First, he notes that empirical researchers can appropriate the definition of a classic work on the concept, which in the case of representation has been the case with Pitkin’s framework. Second, scholars can adopt a “causal-explanatory” understanding of the concept by which he means that a concept can be described by what explains it. For example, in the case of representation, one can say that to represent and be represented is one of the key aspects of Democratic systems, and thus one can only examine representation in terms of how the represented and the one representing interact (Chapter 5 examines this interaction from the viewpoints of Latino legislators). Gerring proposes a schema that focuses on “specific definitional attributes” of the given concept (Gerring 2001). Researchers should formulate a minimal definition of the concept as well as an ideal-type definition. Case Selection Once one performs adequate conceptualization and typology formation, one of the most challenging tasks is to ensure that cases are chosen carefully. Very little agreement exists on how to choose cases.13 As Gerring (2001) points out, one of the main goals of social science is achieving representativeness in a given sample. That is, one should avoid bias in the cases we choose. With respect to the concept of representation, then, it is crucial to choose cases that are representative of the type of representation we are trying to study. For example, it would be unrepresentative to generalize about Mexican American members of Congress by only selecting Rep. Henry Bonilla (R-Texas) for an interview because he was the only Mexican Republican in Congress at the time and his views are not representative of many Mexican Americans.14 Gerring suggests that random samples are always solutions to selection bias, but in the case of racial representation in Congress, it would be foolish to pick a random sample of a population of twenty members of Congress. As the methodology section later describes, I have attempted to select Latinos from various backgrounds, subethnic groups, regions, and genders in my interviews. One of the cardinal sins of social science is selection bias.15 Selection bias 13

14

15

Van Evera offers a list of eleven criteria for case selection, with a matrix aimed at making criteria for case selection easier for graduate students writing their dissertations (p. 88). Please indulge the unfortunate relationship of the concept of representation with representativeness. See Achen and Snidal (1993) who refer to selection bias as an “inferential felony with devastating implications.”

14

Latino Representation in State Houses

occurs when selection takes place on the dependent variable, thus biasing conclusions in many instances. Because qualitative research is more susceptible to selection bias, special attention must be given to methodology in order to avoid criticism that is often unjustified. With respect to studies on representation, Richard Fenno’s seminal work on members of Congress has formed the basis by which other studies of representation have proceeded. His research design involved carefully selecting cases to ensure regional and political representativeness. He did not merely interview whomever he had the opportunity to interview and observe. He carefully selected subjects based on pre-ordained criteria in order to ensure representativeness (more on how I selected my interviews in Chapter 5). He did not assume that more and more interviews would be the answer to selection bias. Had he only chosen Democrats, his study would have been biased based on party affiliation. To assess whether the representatives are representing the represented, it is important to say a few words about what Latino interests are and how, despite the subethnic differences outlined above, there are nevertheless various policies and opinions that unite Latinos. Therefore, the next section turns to what precisely constitutes a “Latino interest.” Latino Interests As the previous discussion on subethnic differences reveals, precisely what constitutes a “Latino interest” varies across time and group. Cuban Americans, Mexican Americans, and Puerto Ricans differ in the importance they place on certain issues; however, on a host of issues, all three groups converge in terms of public opinion. For example, all three groups are generally socially conservative and economically liberal. Conventional wisdom suggests that Cuban Americans are more socially conservative than Puerto Ricans, but as numerous public opinion studies have shown, Puerto Ricans are actually more anti-abortion than Cuban Americans (Uhlaner and Garcia 2002). Although the Republican Party has been trying to raise the salience of abortion among Latino communities, neither of the three major groups as a whole has demonstrated a willingness to vote for the Republican Party based on this position. Many evangelical Latinos, however, have voted for Republicans largely based on their social conservatism. Unlike other Americans, Latinos of all ethnicities generally believe in income redistribution. According to a Pew Hispanic Center and Henry J. Kaiser Family Foundation survey of Latino voters, 58 percent of Latino

Introduction

15

Democrats and 52 percent of Latino Republicans would rather pay higher taxes to support a larger government. Only 35 percent of whites and a surprisingly low 43 percent of African Americans would willingly pay higher taxes to support a larger government. English-dominant Latinos are the least likely (52 percent) to support a larger government, whereas Spanish-dominant and bilingual Latinos strongly support a larger government (62 and 63 percent, respectively). Latinos are also more likely than African Americans and whites to believe that Latino immigrants need to learn English to succeed in America (89 percent versus 86 percent for both African Americans and whites).16 Latinos are also more concerned about certain sets of issues than other groups in American society. In particular, Latinos identify education, health care, economic security, and crime as more pressing issues than whites or African Americans (Griffin and Newman 2008). Because of these issue concerns, Latino legislators should be expected to emphasize their positions on these issues in their three major activities: advertising, credit claiming, and position taking (Mayhew 1974). Indeed, as later chapters will show, Latino legislators pointed to many of these issue concerns when discussing their roles as legislators. Brief Discussion of Subsequent Chapters Chapter 1 introduces the landscape of Latino political representation in the United States, including the historical and theoretical antecedents of this growing demographic. Although some work has focused on the election of Latinos to Congress, little or no attention has been paid to the increasing numbers of Latinos serving in state legislatures. In that chapter, I identify the key arguments in the literature on black, women, and Latino representation and argue that a more nuanced understanding is needed for the study of Latino representation in U.S. legislatures. Namely, the literature has overemphasized roll call voting analyses and ignored the institutional and demographic conditions under which Latinos are elected to legislative bodies. Through an examination of state-level and district-level data, I argue that institutional criteria such as professionalization, term limits, and legislative design have a profound affect on 16

See Pew Hispanic Center/Kaiser Family Foundation National Survey of Latinos: The Latino Electorate, October 2002 for more details. Differences with respect to social and foreign policy issues exist within the Latino community, and this survey reveals such divisions.

16

Latino Representation in State Houses

Latino representation. Moreover, key differences emerge among the key states in terms of the conduciveness of the election of Latinos to certain legislative bodies. Latinos are increasingly getting elected to districts without Latino majorities through a combination of strategic decision making, exploitation of the electoral system, and elite-level assistance. I also make the case that Latino legislators see themselves as distinctive and necessary for the advancement of Latino interests. Chapter 2 provides a comprehensive look at Latino representation in all fifty states by tracking the growth of Latinos serving in legislatures in the past two decades and specifying the extent to which institutional, political, and ethnic factors contribute to Latino descriptive representation. I estimate the extent to which the percentage turnover in each state legislature, the presence of term limits legislation, lower or upper chamber, the Democratic share of the presidential vote, and the level of professionalization have an effect on Latino representation in state legislatures. At the state level, Latino representation in the legislature is largely reflective of the percentage of Latinos in the population. States with larger proportions of liberal voters are also more likely to have higher levels of Latino representation because of elite-driven ways to ensure it, and because liberal voters are more likely to support Latino candidates, who are disproportionately Democratic. Additionally, some states have much higher turnover rates in their legislatures, thus opening up more seats for talented Latinos to win. Alternatively, Latinos may have better chances of being elected in states that have enacted term limits, which eliminates the incumbency advantage. Chapter 3 argues that the Latino presence in state legislatures is not merely a result of demographics, but the percentage of Latino citizens in a district, as well as the institutional design of the state legislature, has important, measurable effects on Latino representation. I use statistical analysis based on my collection of state-level legislative data from seven states going back to the 1990s to determine the extent to which the percentage of Latinos and Latino citizens in a district helps explain the presence of Latinos in state legislatures and Congress. I also test the hypothesis that the percentage of African Americans in a district is positively associated with the presence of Latinos in legislatures, as well as the contention that a Latino candidate is more likely to be elected to the lower-level chamber than to the upper chamber. Chapter 4 explores the various entry opportunities of Latino candidates in non-Latino districts. Using elite interviews and archival research,

Introduction

17

I have compiled the first-ever database of all Latino legislators elected to non-Latino majority districts. This chapter identifies patterns that can help explain the election of the sixty-five Latinos who serve nationwide in non-Latino majority districts. First, I identify elite-driven methods, such as redistricting, recruitment by party leaders, and appointments by governors that have helped enhance the descriptive representation of Latinos in non-Latino majority districts. Second, I examine how several features of the electoral system, such as multicandidate primaries and minority primary coalitions, have also contributed to Latino victories in these districts. Finally, I examine the growing number of Latinos with non-Latino names who have won in non-Latino majority districts and determine the extent to which this has contributed to their success. Chapter 5 argues that Latino legislators see themselves in distinct ways with varying political backgrounds and perceptions of the importance of descriptive representation on substantive representation. I examine how Latino legislators see themselves in terms of political background, their legislative districts, election to current position, political ambition, and their perceptions of representation. This chapter is the culmination of over twenty in-depth, face-to-face interviews with Latino legislators from places as diverse as Florida and Utah. In it, I identify labor unions as common places of early political involvement among Latinos. The chapter also shows that most Latinos identify their districts in racial and class terms, most have considered or are actively considering running for higher office, and most identify different issues, ranging from water rights to stopping gang graffiti in their districts, as highly important to them. As Tip O’Neill once said, “All politics is local.” Although this is true for Latino representatives, most Latino representatives nevertheless indicate a sense of linked fate and indicate that being Latino helps them represent their Latino constituents in a way that non-Latinos cannot. Chapter 6 examines roll call votes in the U.S. House, as well as a select group of state legislatures, to determine if Latino legislators vote differently than their non-Latino counterparts. Through the use of existing Poole-Rosenthal ideology scores, as well as new ideology scores developed for state legislatures, this chapter explores whether having Latinos in high places makes any difference in terms of substantive policy outcomes. Key explanatory variables helping explain ideology include the percentage of Latinos in a district, the percentage of African Americans in a district, whether a district is represented by a Latino, and the political party of the member. While Espino (2003) and Lublin (1997) have

18

Latino Representation in State Houses

examined similar questions, this chapter adds several state legislatures to examine whether the patterns we have seen at the federal level are similar to state-level voting. Chapter 7 returns to the central questions posed at the start of this chapter and provides an overall assessment of the causes and consequences of having Latino legislators. This chapter also assesses the public policy implications of redistricting and term limits on Latino representation in Congress and state legislatures.

1 Latinos in Legislatures Historical and Theoretical Setting

Before the Voting Rights Act in 1965, few Latinos served in the U.S. Congress. Before 1912, only one Latino, California Republican Romualdo Pacheco, served in the U.S. House (Vigil 1996). With the exception of New Mexico and Louisiana, no state sent a Latino to Congress between 1912 and 1960 (Lublin 1997). The Congressional Hispanic Caucus began in 1976 through the efforts of Democratic Reps. Herman Badillo (NY), Baltasar Corrada (PR), E. Kika de la Garza (TX), Henry B. González (TX), and Edward Roybal (CA). Compared to the Congressional Black Caucus, the Hispanic caucus is newer and smaller. In 1992 and 1994, African-American representatives numbered thirty-eight whereas Latinos numbered seventeen (Lublin 1997). In 2005, the number of Latinos serving in Congress had increased to twenty-five. More groundbreaking, however, was the election of two Latinos to the U.S. Senate in 2004. By 2008, Sen. Robert Menéndez (D-NJ) was elected to the U.S. Senate, giving Latinos three U.S. senators for the first time in history. Until 2004, no Latino had served in the Senate since New Mexico’s Joseph Montoya, who served until 1977. As shown in Table 1.1 (both parts a and b), the number of Latinos in Congress had increased to twenty-five by 2008. As shown in Table 1.2, the growth of Latinos in Congress is partly a consequence of the national growth in the Latino population. It especially reflects the concentration of Latinos in California, Texas, and Florida, which now send sixteen Latino representatives to the House (seven for California, six for Texas, and three for Florida). The growth of Latinos in Congress can also be attributed to the redistricting process, which created more districts

19

Latino Representation in State Houses

20

Table 1.1a. Latino Representatives in the U.S. Congress Ranked in Descending Order of Latino Population, 2009 Name

District

Rubén Hinojosa Henry Cuellar Silvestre Reyes Lucille Roybal-Allard Luis V.Gutiérrez Grace Napolitano Lincoln Díaz –Balart Xavier Becerra Solomon Ortíz Charles González Loretta Sánchez Ciro Rodríguez José Serrano Ed Pastor Ileana Ros-Lehtinen Mario Díaz-Balart Hilda Solís Linda Sánchez Joe Baca Raúl Grijalva Nydia Velázquez Albio Sires John Salazar Bob Menéndez

15-Texas 28-Texas 16-Texas 34-California 4-Illinois 38-California 21-Florida 31-California 27-Texas 20-Texas 47-California 23-Texas 16-New York 2-Arizona 18-Florida 25-Florida 32-California 39-California 43-California 7-Arizona 12-New York 13-New Jersey 3-Colorado Sen-New Jersey

% Latino

% Black

78 78 78 77 75 71 70 70 68 67 65 65 63 63 63 62 62 61 58 51 49 48 22 13

2 1 3 4 4 4 7 4 3 7 2 3 36 5 6 10 3 6 3 3 9 11 1 13

Party Democrat Democrat Democrat Democrat Democrat Democrat Republican Democrat Democrat Democrat Democrat Democrat Democrat Democrat Republican Republican Democrat Democrat Democrat Democrat Democrat Democrat Democrat Democrat

Source: National Journal’s Almanac of American Politics, 2009.

with significant numbers of Latinos.1 The National Association of Latino Elected and Appointed Officials (NALEO) estimates that 122 (28 percent) of the 435 U.S. House districts have Latino populations that surpass the national average.2 The Southwest and California clearly have the highest percentages, but Latinos have surpassed African Americans as the largest minority group in other states, including New Jersey and New York. This book will explore in more detail the determinants of the growth in the number of Latinos serving in Congress and many state legislatures. 1

2

Ex-Rep. Bob Menéndez (D-NJ) won his seat in 1992, following redistricting, whereas Rep. Mario Díaz-Balart (R-FL) won his seat in 2002 following redistricting. Obtained from NALEO Election Guide, published and available at http://www.naleo. org, 2008.

Latinos in Legislatures

21

Table 1.1b. Non-Latino Representatives in the U.S. Congress Ranked in Descending Order of Latino Population (40 Percent and Above Districts), 2009 Name

District

Gene Green Jim Costa Howard Berman Bob Filner Sam Farr Charlie Rangel Maxine Waters Harry Teague Devin Nunes Laura Richardson Martin Heinrich Dennis Cardoza Lois Capps

29th District-TX 20th District-CA 28th District-CA 51st District-CA 17th District-CA 15th District-NY 35th District-CA 2nd District-NM 21st District-CA 37th District-CA 1st District-NM 18th District-CA 22nd District-CA

% Latino

% Black

66.1 63.1 55.6 53.3 49.2 47.9 47.4 47.3 43.4 43.2 42.6 41.9 41.7

9.7 7.2 4.1 9.4 2.6 30.5 34.1 1.6 2.1 24.8 2.3 5.6 1.9

Party Democrat Democrat Democrat Democrat Democrat Democrat Democrat Democrat Republican Democrat Democrat Democrat Democrat

Source: National Journal’s Almanac of American Politics, 2009.

Table 1.2. Percentage Latino in Seven States and United States, 1950–2000 (Census Bureau)

New Mexico California Texas Arizona Florida New York New Jersey United States of America

1950

1960

1970

1980

1990

2000

2004

– – – – – – – 1.5

– – – – – – – 2

37 14 18 17 7 8 4 5

37 19 21 16 9 10 7 6

38 26 26 19 12 12 10 9

42 32 32 25 17 15 13 13

42 34 34 25 18 16 14 14

In the fifty state legislatures, most Latino representatives have been elected from majority-Latino districts, but there are also some Latino legislators who have been elected from districts without Latino majorities. In 2004, 157 Latino legislators represent majority-Latino districts, including 132 Latino legislators from the seven states analyzed in this book (New Mexico, California, Texas, Arizona, Florida, New York, and New Jersey).3 Sixty-five Latinos currently serve non-Latino majority districts. 3

A rationale for these seven states will be discussed later in the book.

Latino Representation in State Houses

22

Table 1.3. Percentage Latino in Seven State Legislatures and Congress, 1986–2002

New Mexico California Texas Arizona Florida New York New Jersey US Congress

1986

1990

1994

1998

2002

2004

30 6 13 13 5 3 1 2

34 5 15 13 7 3 1 2

38 10 19 11 9 5 1 2

35 15 20 10 9 6 3 3

39 22 19 18 9 6 4 4

40 24 20 18 10 7 5 5

Source: Percentages computed based on National Directory of Latino Elected Officials (published by the National Association of Latino Elected and Appointed Officials), 1986–present.

Table 1.3 shows the growth of Latinos serving in Congress and the seven legislatures since 1986. The most pronounced growth since 1986 has been in California, where the percentage of Latinos serving in the Legislature increased from 6 percent in 1986 to 24 percent in 2005. In all seven states, the percentage of Latinos serving in legislatures has increased substantially since 1986. In 1986, Latinos comprised 30 percent of the New Mexico Legislature. Latinos now comprise 42 percent of the New Mexico Legislature, slightly higher than the corresponding percentage of Latinos in the state population. The reasons why New Mexico is so different are explored in Chapter 3. Some states, such as New York, have experienced growth in the percentage of Latinos serving in the Legislature, but not nearly as significant as the growth in California or Texas. The reasons for these differences are also explored in Chapter 3. Literature on Latinos in Legislatures The literature on Latino representation in the U.S. Congress is quite sparse, although political scientists have conducted several studies in recent years. The earliest work on Latino representation borrowed heavily from previous work on African-American representation in Congress. Table 1.4 lists some of the major works on race and representation since 1984, and most deal with African-American representation in the U.S. Congress. This table describes the key research question of the work, the dependent variable (s), and the methodology used. Note that statistical analyses are by far the most utilized method, indicating the prevalence of

Table 1.4. Literature on Racial and Ethnic Representation in U.S. State Legislatures Author (s)

Research Question

Dependent Variable(s)

Methodology

Welch & Hibbing (1984)

Do Latino members of Congress vote differently than non-Latino members? Do black representatives better represent black constituents? Are Latinos substantively represented?

Conservative coalition scores

Regression Analyses

LCCR scores

Regression analyses and case studies Regression analyses

Swain (1993) Hero & Tolbert (1995)

Kerr and Miller (1997)

Are Latinos substantively represented?

Whitby (1997)

What is the link between black populations and House member behavior? Are majority-minority districts needed to elect minority legislators? Do black majority districts promote politics of difference?

Lublin (1997)

Canon (1999)

Tate (1999)

Bratton (2006) 23

Do blacks need to be descriptively represented in Congress to be well represented? What is the effect of the Latino population on state legislative behavior?

Southwest Voter Research Initiative scores for 100th congress Southwest Voter Research Initiative scores for 100th congress LCCR and other key votes

Regression analyses

Regression analyses/ logistic and OLS

DW-NOMINATE scores; percent minority in district

Regression analyses/ logistic and OLS

Cosponsored bills, speeches, LCCR scores

Regression analyses, content analyses, case studies, interviews Regression analyses

Votes on major legislation, NBES for black constituent attitudes Committee assignments, bill passage, and bill sponsorship

Regression analyses/negative binomial and logistic

24

Latino Representation in State Houses

roll call voting analyses. In the next several paragraphs, these works are discussed in more detail. Many scholars assumed Latinos were a monolithic group, generally more liberal than whites were and much more similar in terms of political behavior to African Americans. Welch and Hibbing (1984) noted that Latino Conservative Coalition scores were more liberal than non-Latino representative scores. This study, however, only examined members from 1973 to 1980. It was not until 1992 that members of Mexican, Puerto Rican, and Cuban groups were simultaneously represented in Congress. Earlier Latino representatives were mostly Mexican Americans who are one of the more liberal Latino groups in the United States. Hero and Tolbert (1995) went beyond Welch and Hibbing’s (1984) earlier analysis by using Southwest Voter Research Initiative (SWVRI) scores for the 100th Congress to gauge the representation of Latinos and their interests. In their analysis, Hero and Tolbert found that high SWVRI scores for Latino representatives were not significantly different from non-Latino representatives.4 In essence, they found that Latinos might benefit from “collective representation” and that dyadic representation was not evident. Kerr and Miller (1997) responded to this article by arguing that dyadic representation of Latino interests was in fact present. For them, “dyadic and collective representation can and do occur simultaneously in the political system and, as an analytical matter, should be considered together” (Kerr and Miller 1997, 1071). Although exposing some methodological problems with the paper, Kerr and Miller do not provide the necessary prescriptions for a better analysis. For example, the SWVRI scores in question are few in number and only cover the 100th Congress. Updated data reflecting the latest increases in the Latino population are needed for a more comprehensive analysis, as Chapter 6 will demonstrate. David Lublin (1997) uses Poole-Rosenthal NOMINATE scores as the dependent variable for his analyses. Poole-Rosenthal scores do a much better job of assessing the political ideology of members of Congress because they include all votes, not just a few select votes, such as Americans for Democratic Action, American Conservative Union, or Southwest Voter Research Institute scores. The Poole-Rosenthal scores are continuously distributed, unlike interest group ratings, which are lumpy (Lublin 1997). 4

SWVRI scores measure the extent to which legislators voted in support of a “Latino agenda” as determined by an umbrella group of Latino organizations. High scores indicate legislative voting records with support of Latino issues.

Latinos in Legislatures

25

Lublin interacts the percentage of Latino population with the party of the representative and finds that Republican members are more conservative when they have higher Latino populations, whereas Democratic legislators are significantly more liberal (Canon 1999a; Lublin 1997b). Lublin explains this by noting that Democratic legislators tend to represent Mexican Americans and Puerto Ricans whereas Republican legislators generally represent Cuban Americans. Another, more recent study examining Latino representation in Congress is Espino’s (2003) work analyzing roll call voting data from the 103rd to 107th Congresses. He finds little support for the contention that members of the Congressional Hispanic Caucus vote more often with their fellow caucus members than they do with other members from their state delegations or with legislators who share their ideologies. Another significant drawback of previous research on Latino representation is that it has focused almost exclusively on the U.S. Congress. Whereas this book does explore the election of Latinos to Congress, the value-added analyses will focus on the much-neglected state legislatures (regarding Latino representation). Additionally, previous research has focused heavily on the extent of Latino substantive representation without first answering some basic questions about descriptive representation. For example, Kathleen Bratton’s (2006) recent study demonstrates that in seven U.S. legislatures, the presence of Latino legislators and Latino constituents has important implications for legislative behavior and success. Few scholars have inquired about what kind of districts tend to elect Latino legislators and analyzed how a district’s ethnic composition affects the probability of electing Latino legislators (Grofman and Handley 2001; Lublin 1997). Moreover, few have asked how Latino legislators see themselves and the issues they care about in their districts. Chapter 5 will argue that Latino legislators see themselves in distinctive ways as demonstrated by a series of interview questions conducted with more than twenty legislators from diverse backgrounds and districts. Other Minorities in Legislatures and Redistricting The literature on African Americans and women in legislatures has provided unique insights into how minorities have been elected to Congress, which may be applicable to Latinos but not entirely appropriate. Most of the research regarding African Americans in legislatures has tended to focus on the U.S. Congress. In the area of racial representation, Carol Swain’s (1993) Black Faces, Black Interests is an analysis of how

26

Latino Representation in State Houses

African Americans are represented in Congress. Swain’s interviews with African-American members of Congress provide great insight into the varying styles within the African-American community. Her analyses of former Rep. Mike Espy (D-MS) and Rep. John Conyers (D-MI) show just how differently African-American members of Congress responded to their constituencies in order to secure re-election. In a similar way, Rep. Lincoln Díaz-Balart (R-FL) and Rep. Xavier Becerra (D-CA) represent their districts in different ways, which is more often the case with Latinos who have a much wider range of experiences. Swain’s fundamental thesis that blacks would be better served by electing Democratic members of Congress, regardless of race, has been a controversial one. For example, Canon (1999a) notes that Swain does not account for all white representatives in districts with at least a 25-percent African-American population in the 103rd Congress. Because Latinos are not as solidly Democratic as African Americans, it does not follow that merely electing more Democrats to Congress would necessarily be a positive thing for the Latino community. On the contrary, political parties are increasingly courting Latino voters precisely because of their willingness to vote in less predictable ways. In the case of Latinos, we have definitely seen more biracial coalitions but not always with African Americans (Casellas 2009b). In his book, Race, Redistricting, and Representation, David Canon (1999a) defends what he terms a “supply-side” theory of racial redistricting. He argues that the type of racial representation that exists in a district is dependent on the racial composition of contenders in the Democratic primary. In his view, the demand-side of the equation (voters) is less instrumental than the supply-side (candidates). Thus, at times, Latinos join forces with whites against the African-American community’s favored candidate. Mayor Richard M. Daley (D-Chicago), for example, was initially elected with a Latino-white coalition. This coalition appears in many legislative districts in Democratic primaries. Recent studies regarding African Americans and Latinos in legislatures indicate that the growing numbers of minorities in legislatures have resulted in the advancement of minority interests (Haynie 2001; Preuhs and Hero 2009). The debate regarding the value of descriptive representation for yielding substantive representation is an important one that has been debated by Mansbridge (1999), Canon (1999b), and Whitby (1997), who place a greater emphasis on descriptive representation, whereas Hero and Tolbert (1995) and Swain (1993) place a greater emphasis on substantive representation. Most studies on representation

Latinos in Legislatures

27

take for granted what we know about descriptive representation and leap to questions about substantive representation. This book fills the gap in the descriptive representation literature by arguing that simply assuming that demographic growth is the only way for increasing Latino representation misses the key institutional and state-level differences that help explain the election of Latino candidates. It also fills the gaps in the Latino representation literature by expanding our knowledge of the election of Latinos to state legislatures, with rich insights gained from the collection of large state- and district-level data, participant observation, and interviews with more than twenty Latino legislators from all parts of the country. In addition, research previously done on African-American representation must be updated in light of the differences between both minority groups and the contributions of scholars of legislative politics. Previous studies on racial representation have not been as successful as they could be in synthesizing findings about race and ethnicity with the more general findings about legislative elections and legislative behavior. Studies of women in legislatures have been plentiful, and we know that Latinas have different electoral strategies and outcomes in legislative races than Latinos. Copeland et al (2004) find that women legislators tend to come from more liberal, northern, and better educated states. Perhaps a more important finding in their study is that the supply of candidates is the one variable common to the success of both women and African Americans. Burrell (1990) makes the important point that the presence of women in legislative institutions has important agendasetting implications. Fraga and Navarro (2004) have found that Latinas in Texas tend to identify more with women than with Latinas, whereas the opposite is true in California, yet at the same time, Latinas in general are more liberal than Latinos are and less likely to support increases in defense spending. In this sense, Latinas are no different from women writ large (Bratton and Haynie 1999). Latinas have also been more successful winning U.S. House races in California than in Texas. Additionally, states with multimember districts, like Arizona and New Jersey, tend to have higher percentages of Latinas serving in legislatures. The literature on African Americans and Latinos in legislatures, however, is incomplete in that studies to date have not adequately addressed the conditions under which Latinos are elected to Congress and state legislatures in the first place. We know about how Latino members of Congress vote (Espino 2003; Lublin 1997), but not the conditions under which they are elected to their districts. Additionally, we know

28

Latino Representation in State Houses

very little about the conditions under which Latinos are elected to state legislatures – the very bodies in which many Latino members of Congress previously served. Too much of the existing literature focuses on roll call voting in Congress, with insufficient attention devoted to how Latino legislators are elected to legislative bodies, how Latino legislators view themselves, and the effects of institutional structures and public policies such as term limits on Latino representation. Yet another limitation in the existing literature is that to date, no study compares Latino roll call voting in Congress with roll call voting in state legislatures.5 Research Questions Some of the central questions driving this book are as follows: (1) Under what conditions are Latinos elected to Congress and to state legislatures? Although it is true that many Latinos are elected from majority-Latino districts, substantial numbers of Latino legislators do not represent districts with majority-Latino populations. (2) How much does the ethnic composition of a district affect the chances that a Latino candidate will get elected to office? Challengers to incumbents often face an uphill climb in their efforts to be elected to a given office. Latino candidates may find it advantageous to run in a district with a majority-minority population. However, only one Latino in the U.S. House represents a majority-white district. (3) How much does the ethnic composition of a district affect the chances that a Latino candidate will be elected in a given district? At least at the national level, very little evidence exists that Latinos are being elected from districts with white majorities. To date, only one district with a combined white and Latino population majority has elected a Latino member of Congress. At the state level, this pattern differs, which is why it is crucial to understand the variables that contribute to Latino victories at the state level. I argue that Latinos are primarily elected from districts with substantial majorities of Latino citizens, much in the same pattern as African Americans. However, Latinos can be assisted by institutional characteristics, such as the type of legislature. The probability of electing Latinos to the legislature will vary by state, depending on a host of variables. Some of the explanatory variables that help explain the election of Latino 5

We must, of course, be careful about direct comparisons between Congress and state legislatures, because the type of legislation and issues do vary within and across legislative bodies.

Latinos in Legislatures

29

legislators are district demographics (percentage Latino in district and percentage African American in district) and the belief that Latinos will have more difficulty getting elected to the upper chamber. Latino candidates are more likely to run in districts with higher percentages of Latinos because their chances of winning would disproportionately increase (See Table 1.1 for Latino U.S. House representatives). I would expect these variables to have an impact on Latino descriptive representation in legislatures based on previous scholarship on Congressional elections. Latinos are not unlike other candidates for legislative positions in the considerations that they employ to decide whether to run for office. Certainly, if an incumbent retires because of term limits and a seat becomes available, more candidates may decide to run for political office.6 Incumbency no doubt confers enormous benefits, and challengers decide to run based on whether an incumbent is in the race. The ability to raise money matters too, but it is more critical for challengers than incumbents.7 Incumbents decide to run in districts that will give them the highest probability of winning. Representatives often come from districts with which they share similar demographic characteristics. Conversely, representatives choose to run in such districts. This is most clearly reflected in the way representatives interact with their constituents at home (Fenno 1978). Much in the same way Swain (1993) demonstrates that African Americans are more at ease with fellow African-American legislators, Latinos are also more likely to trust Latino representatives (Pantoja and Segura 2003). There is no question, then, that districts with higher proportions of Latinos will be more likely to support Latino candidates. After explaining the conditions by which Latinos are elected to state legislatures and Congress, the key questions become: (1) What difference does it make whether Latinos are elected to Congress and state legislatures? By virtue of being Latino, legislators may care about different issues or more strongly advocate positions on a subset of issues, such as education policy. (2) To what extent do Latino legislators value descriptive representation? (3) More pointedly, do Latino legislators vote differently from others who represent similar constituencies? Studies of Latino public opinion have shown that Latinos tend to be fiscally liberal 6

7

See Fowler and McClure (1990) for more on the dynamics of potential candidates in congressional districts, as well as the role of political ambition in determining who runs for Congress. See Cox and Katz (1996) and Jacobson (1997) for more on the incumbency advantage and factors determining who runs for congressional districts.

30

Latino Representation in State Houses

and socially conservative.8 Many Latinos in Congress and state legislatures, however, appear to side with the Democratic Party on both fiscal and social issues. This section is an attempt to ascertain the extent to which Latinos differ from other representatives in terms of voting behavior. (4) Do Latino legislators emphasize different types of interests than other legislators who represent similar constituencies? Being Latino may explain why some legislators fight for causes not otherwise important to other groups. For example, Cuban American legislators may have a much more passionate view about the embargo on Cuba than other Latinos because of their personal experiences. Likewise, Mexican Americans will be more sensitive to immigration policy than Puerto Ricans because the policy more directly affects them and their immediate community. Types of Data I have collected comprehensive information on Latino elected officials in all seven of the state legislatures, as well as Congress. This will enable me to test the changes from 1990 to 2000 in terms of redistricting and the effect on Latino representation. The National Association of Latino Elected and Appointed Officials (NALEO) has identified Latino legislators for each chamber in the fifty states in an annual publication from 1984 to the present. The United States Census Bureau collected all the demographic data used in this study. The allocation of demographic data to legislative districts was done by each state and posted on statemaintained websites. The demographic variables are the percentage of a given district’s population identified in the most recent census as Latino or African American. I have assembled additional demographic data for both current legislative districts and for districts before the 2001 redistricting cycle. Using elite interviews and archival research, I have assembled the firstever database of Latino legislators elected to non-Latino majority districts in all fifty states. As before, the identification of Latino legislators in each chamber was done by the NALEO. This section of the book, however, is not limited to the seven states. In fact, thirty-six of the sixty-five Latinos serving in non-Latino majority districts come from the other forty-three 8

See Kaiser Foundation study for more details. The majority of Hispanics of all backgrounds (Cubans, Mexican Americans, and Puerto Ricans) believe in a larger role for government yet are opposed to abortion and gay rights legislation.

Latinos in Legislatures

31

states. To determine whether a Latino represents a non-Latino majority district, I used Census Bureau data to determine the percentage of Latinos and African Americans in each district and compiled a database including this information. I also collected additional information on the district and the member, such as the political party of the member, the subethnic group of the member, the chamber of the member, whether the member was first elected in an open seat, the year first elected to any political office, and the year elected to the current office. This information was collected using Lexis-Nexis, candidate websites, state legislature websites, and phone calls to legislative staff members. In the past several years, I have travelled to the annual meetings of the NALEO and the NHCSL. I have gleaned insights from participant observation in these meetings, which were heavily attended by Latino legislators from nearly all parts of the country. I also conducted in-depth interviews with more than twenty Latino legislators from diverse backgrounds (see Table 5.1 for descriptive statistics). The bulk of Chapter 5’s data and analysis stem from these interviews and participant observation at these meetings. I also use parts of the dataset assembled by David Lublin in his analysis of Congressional roll call voting. His dataset includes PooleRosenthal scores and additional demographic variables for U.S. House districts from the 87th through the 101st Congress.9 I have added PooleRosenthal DW-NOMINATE scores from the 102nd Congress through the 104th Congress to Lublin’s dataset.10 Important variables to notice in this dataset include dummy variables for Cuban, Mexican, and Puerto Rican representatives.11 This dataset is the most comprehensive available for studying Latino representation at the national level. I also use a dataset assembled by Nolan McCarty for analysis of state legislative roll call voting. His dataset ascribes Poole-Rosenthal NOMINATE scores to members of most state legislatures in 1999 and 2001 for Colorado, New Jersey, and Texas. I have merged his dataset about roll call voting with my dataset about candidate ethnicity and district demographics to determine the impact of Latino districts and Latino legislators on legislative behavior.

9 10

11

Data analyses conducted using STATA statistical software (Intercooled Stata 7.0). DW-NOMINATE scores are used in this analysis because such scores are comparable within and across Congresses. To date, all Latinos elected to the House of Representatives have been Cuban, Mexican, or Puerto Rican.

32

Latino Representation in State Houses

Choice of Legislatures In addition to Congress, the seven state legislatures that are included in Chapter 3 of this study are New Mexico, California, Texas, Arizo`na, Florida, New York, and New Jersey. I chose these seven states and Congress because of their geographic and subethnic-group diversity in order to rule out region-specific or subethnic-group anomalies. These are the states with the largest percentage of Latinos, but states like Nevada have recently experienced a surging Latino population and could very well have been included in this study. Random selection of states does not work because some states have a tiny Latino population and no Latino legislators, thus making analysis impossible. These states all have significant and growing Latino populations. New Mexico is an obvious choice because of its 42 percent Latino population and equally high percentage of Latinos in its legislature. California is the only majority-minority state. It has a 34 percent Latino population, and its legislature is one-quarter Latino. Like California, Texas has a 34 percent Latino population, but the percentage of Latinos in the Texas Legislature is 20 percent. Latinos comprise one-quarter of Arizona’s population, and 18 percent of the Legislature is Latino. Florida’s Latinos number 18 percent of the population, whereas 9 percent of the Legislature is Latino. In New York, Latinos comprise 16 percent of the population but only 6 percent of the Legislature. In New Jersey, Latinos comprise 14 percent of the state, whereas only 4 percent of the state legislators are Latino.12 The next chapter assesses the effects of turnover, term limits, and other institutional and demographic determinants of Latino representation in all state legislatures. 12

Percentage Latino in state obtained from NALEO 2004 Election Handbook (see http:// www.naleo.org), whereas percentage Latino in legislatures obtained from my own research (see Chapter 3).

2 The Effects of Population, Turnover, and Term Limits on Latino Representation

To what extent has the growing Latino population translated into more Latino legislators in Congress and state legislatures? Are Latinos as likely to serve in upper chambers as lower chambers? Are Latino candidates advantaged when turnover is high within legislatures? Does the imposition of term limits further advantage Latino candidates? Are states that are more liberal more likely than conservative states to elect Latino candidates? This chapter examines Latino representation in all fifty states by tracking the growth of Latinos serving in legislatures over the past decade and comparing the percentage of Latino citizens in the population with the percentage of Latinos in each legislature. At the state level, the Latino representation in the legislature is largely reflective of the percentage of Latinos in the population. Additionally, some states have much higher turnover rates in their legislatures, thus opening up more seats for talented Latinos to win. Similarly, Latinos may have better chances of being elected in states that have enacted term limits, which eliminates the incumbency advantage. Using the latest data on turnover and term limits, I help explain the impact of these variables on Latino representation in legislatures (Moncrief, Niemi, and Powell 2004). Chapter 3 explores the conditions under which Latinos are elected to seven of the nine state legislatures with the largest Latino memberships, and will provide precise estimates of district-level effects. Latino Representation in State Lower and Upper Chambers The pattern for most states is that Latino descriptive representation lags, due to many factors, including historically low rates of voter turnout,

33

34

Latino Representation in State Houses

a large number of undocumented immigrants, and higher rates of noncitizenship among Latinos. At the same time, we find some variation between lower and upper chambers across the state legislatures. As Figures 2.1 and 2.2 show, by 2000, New Mexico remained an outlier, but California had shown a dramatic increase in Latino representation compared to 1992. Florida’s Latino population has increased in eight years, but Latino legislative representation in the lower chamber has remained stagnant. Nevada’s Latino population has also substantially increased, yet Latino legislative representation has not risen with this phenomenon. The pattern for upper chambers looks very similar to lower chambers, with the exception of a few states, such as Florida and Nevada, which have a much higher percentage of Latinos in the State Senate than in the House. As in the lower chamber, the percentage of Latinos serving in the California Senate increased over the eight-year period. Only in Arizona and Texas did the percentage of Latinos serving in the Senate increase vis-à-vis 1992. In Florida, the Latino population increased from about 12 percent to 17 percent, yet the percentage of Latinos in the State Senate decreased to less than 10 percent of the total. By 2004, we have seen only minor changes from 2000. Colorado had a slight decrease in Latino representation in the House, whereas the Senate exhibited little change (most probably because of the longer terms). Turnover, Professionalization, and Term Limits I expect that states with higher percentages of Latinos in the population would have higher percentages of Latinos in the legislature.1 The percentage of Latinos in both upper and lower chambers is substantially below the corresponding percentage of Latinos in the state population. New Mexico is clearly an outlier in that there is parity between the state population and legislative descriptive representation. Arizona and Texas are the only other states with double-digit representation in their legislatures, despite the fact that six other states had more than a 10 percent Latino population in 1992. Few studies have examined the effect of the Latino population, legislative professionalization, and state legislative term limits on Latino representation. As Moncrief et al. (1996) observe, “[w]hile there has 1

Just as districts with higher proportions of Latinos will be more likely to elect Latino candidates, as the Latino population increases in a state, the percentage of Latinos in the legislature should also increase (Lublin 1997).

Effects of Population, Turnover, and Term Limits

35

40

NM

Per Latino legislator

30

20 TX

10

CO FL

IL CT WA MA KS MN UT IN RI PA WV KY MS ME AL TN ND VT SD AR SC NH NC IA MO OHMD MT GA WI WY MI LA DE VA OK AKORID HI

0

0

AZ

CA

NY NJ NV

10

20

30

40

Per Latino

Figure 2.1. Latino representation in state lower houses, 1992.

NM

40

Per Latino legislator

30

20

TX FL

10

NV KS WA IL WY ID MA WV AL MS KY ME TN NDWI VT SD AR SC NHMI IA NC MO MNVA OH MT GA IN PA LA NE DE MD OAK KOR RI UT CTHI

0 0

AZ CO NY

CA

NJ

10

20

30

Per Latino

Figure 2.2. Latino representation in state upper houses, 1992.

40

Latino Representation in State Houses

36

Table 2.1. Term Limits Legislation in the United States State

Year Enacted

Impact Year

Arizona Arkansas

1992 1992

California

1990

Colorado Florida Louisiana Maine Michigan

1990 1992 1995 1993 1992

Missouri Montana Nebraska Nevada Ohio Oklahoma S. Dakota

1992 1992 2000 1996 1992 1990 1992

2000 1998 House 2000 Senate 1996 House 1998 Senate 1998 2000 2007 1996 1998 House 2002 Senate 2002 2000 2006 2010 2000 2004 2000

Source: National Council of State Legislatures. Excluded are states that passed term limits that were either overturned by the courts or repealed by the legislature.

been research on the consequences of women in state legislatures, similar research on the effect of African-American or Hispanic legislators has been lacking” (310). Broader studies have tended to focus on the impact of term limits on turnover, membership composition, legislative professionalization, and partisanship (Fiorina 1994, 1999; Meinke and Hasecke 2003; Moncrief, Niemi, and Powell 2004). Because many states adopted term limits in the early 1990s, the long-term effects of this policy are only beginning to emerge (see Table 2.1). The literature on women in legislatures has largely focused on the effects of having women in legislatures (Kathlene 1994; Richardson and Freeman 1995; Thomas and Welch 1991). Some studies have examined the effect of reapportionment and the electability of women to legislative institutions, although not explicitly the effect of term limits (Bullock 1992; Grofman and Handley 1989; Nelson 1991; Philpot and Walton 2007). But are Latinos really newcomers to the political system? In most states, this is true, but in states like New Mexico, Latinos have long been

Effects of Population, Turnover, and Term Limits

37

a part of the political process at the elite level (Vigil 1996). In New York and California, the growth of the Latino population has been quite stark in the last decade. This trend is likely to grow, and we are beginning to see the growth of Latino elected officials in unlikely areas of the country, and with the support of white voters rather than just coethnics. As Moncrief et al. (2000) argue, the number of minority legislators would undoubtedly grow if political parties and interest groups made concerted efforts to recruit and fund potential candidates. Fowler (1992) has observed (in the context of women legislators) that the basic problem is that too few women run in most states, either in primaries or in general elections, to produce measurable increases in female representation. This has arguably been the case with Latinos, despite their growing electoral influence. State-Level Diversity Why might higher proportions of Latinos in a state be related to higher proportions of Latino legislators? As Hero and Tolbert (1996) have argued, a state’s racial and ethnic diversity can have important implications for minority groups, especially in homogenous states. The Latino populations in the states are also characterized by differences based on national origin and political incorporation in state politics. Increased diversity at the state level can also create an environment more conducive to the election of minorities. Additionally, such diversity can also have an impact on the election of Latinos to districts that are not ethnically diverse because of the state-level diversity. As Gay (2001) argues, minorities are more likely to trust members of their own group. In addition, Pantoja and Segura (2003) show that as the number of Latinos serving in the legislature increases, levels of political alienation among Latino voters decrease, arguably because of the trust Latinos place in members of their own ethnicity to help solve the problems affecting their community. Barreto, Segura, and Woods (2004) show that Latinos are more likely to turn out when they live in majority-Latino districts. As Bobo and Gilliam (1990) have demonstrated as the “empowerment” hypothesis, minorities become more likely to participate and vote in politics as their numbers in elected positions increase. In particular, when Latinos are represented in large numbers in a state legislature, even white voters become more willing to support Latinos for elective offices. For instance, Bill Richardson (D-NM) was elected governor in a state with a large Latino population (43 percent) precisely because of his ability to appeal to both Latinos and whites in the state.

Latino Representation in State Houses

38

Hence, the first hypothesis is: Demography Hypothesis: The percentage of Latinos in a state’s population will be positively associated with the representation of Latinos in the legislature.

High Rates of Turnover Help Newcomers to the Political Process In recent years, legislative scholars have been interested in explaining the rates of turnover across the states. Turnover rates had declined throughout the 1980s, but the advent of term limits in many states in the 1990s seemed likely to increase turnover rates once again (Niemi and Winsky 1987). Fifteen states currently have term limits. Most term limits legislation was passed in the early 1990s, although there are several key exceptions. Other states, such as Idaho and Oregon, had passed term limits legislation, but the courts or the state legislature proceeded to invalidate them. Of the fifteen states, only a few are states with relatively high percentages of Latinos. Arizona, California, Colorado, and Florida all have significant Latino populations, whereas the other eleven states are generally more conservative with few minorities. According to Breaux and Jewell (1992), the lower turnover rates in the 1980s were due to lower rates of voluntary retirements. It makes sense that states with legislative term limits would see higher levels of turnover and interchamber movement. Indeed, whether a state has a professional legislature, and when term limits were adopted and implemented, all have effects on the turnover percentage (Moncrief, Niemi, and Powell 2004). Although most of the arguments for term limits revolve around getting rid of entrenched incumbents, there is a possibility that this policy has actually helped minorities by opening up seats that would have otherwise not been available to political newcomers. Because Latinos are political newcomers, the effect of term limits and turnover on Latino representation demands scholarly attention. I would expect that states with term limits would have a positive association with the percentage of Latinos in legislatures. That is, legislative term limits may actually increase the ability of Latinos to win in state legislative races. Moreover, related to term limits, I would expect that states with higher percentages of legislative turnover would more likely be associated with higher levels of Latino representation for many of the same reasons. As political neophytes compared to African Americans, Latinos stand to benefit from open seats, especially in states that have seen exponential rates of growth among the Latino community. The little work on the effects of term limits on the representation of minorities and women demonstrate some positive findings for Latino

Effects of Population, Turnover, and Term Limits

39

candidates. Carroll and Jenkins (2001) find that more women were turned out of office by term limits in 1998 and 2000 than were elected to seats open because of term-limited incumbents. They also find that the number of minorities in term-limited state lower chamber seats increased following the elections of 1998 and 2000. Even though they are not specifically concerned with Latinos in their analysis, they stipulate their findings are “somewhat more promising for Latinos who were able to win several term-limited seats in districts where Latinos were a minority (and even a small minority). The findings suggest that Latinos may continue to benefit from term limits if strong Latino candidates who can appeal to primarily white constituencies continue to come forward” (Carroll and Jenkins 2001, 12). As Canon (1999) notes, “states continue to serve as laboratories of experimentation… state-level variation [regarding] … term limits allows some additional leverage on representational questions; research on these topics should help inform debates at the national level” (371). In line with Canon’s call for more research, this chapter seeks to explain the effects of term limits and legislative turnover on Latino representation in the fifty states. That is, term limits may have an unintended positive effect on Latino representation by replacing entrenched incumbents with newcomers to the political system. Work by political scientists in the late 1990s involved the development of models based on past turnover rates and the length of term limits (Francis and Kenney 1997). In 2000, Francis and Kenney proposed a theory of “churning” by which members would leave legislatures for other careers in anticipation of the turnover deadline. These theories no doubt apply to all legislators regardless of race, ethnicity, or gender, but this study seeks to test whether Latinos have been particularly helped or hurt by such policies. Accordingly, Turnover Hypothesis: State legislatures with higher turnover rates will have a higher representation of Latinos in the legislature.

State legislatures are not one-size-fits-all institutions. Scholars have distinguished the level of professionalization of legislatures based on how often they meet, how much legislators are paid, and the size of the staff (Fiorina 1994; Hero 1992). Some state legislatures, like New York’s, are highly professional. That is, legislators are paid a comfortable salary (about $80,000 per year), meet quite often, and have other benefits. Other legislatures, like New Mexico, are known as citizen legislatures. They often do not receive a salary, meet a couple of months during the

40

Latino Representation in State Houses

year, and consequently have higher rates of turnover.2 Fiorina (1994) has hypothesized that professional legislatures are more conducive to the election of Democrats, since potential Democratic candidates tend to come from the public sector and would be attracted by a comfortable position. Although it is true that most Latinos are Democrats, we cannot deduce that Latinos will fare better in professional legislatures for several reasons. First, Latinos are political newcomers, unlike African Americans or labor union candidates. Political newcomers, by definition, have not amassed the networks and name recognition needed to win highly desirable seats. Second, any position that is highly desirable is bound to attract more applicants. Just as we do not see too many Latino CEOs, it is also unlikely for Latinos to be able to compete as successfully for such desirable positions. Institutional Design Matters If we assume that candidates are rational actors who balance the costs and benefits of seeking elective office, we can conclude that more individuals would be attracted to higher-paying legislatures than low-paying legislatures (Mayhew 1974; Downs 1957; Arnold 1990). A perfect example of a professional legislature is New York, which provides a generous compensation of $79,500 per year for members of the legislature, in addition to a “lulu” aimed at covering expenses.3 Because citizen legislatures are less desirable, the pool of candidates should be smaller, which should benefit political newcomers. Latinos are political newcomers, so I would suspect, then, that states with citizen legislatures would more likely be associated with greater levels of Latino representation because turnover is higher and competition is less fierce for these seats, giving disadvantaged groups better opportunities to win.4 This leads to the third hypothesis: Institutional Design Hypothesis: States with professional legislatures will have lower representation of Latinos in the legislature.

2

3

4

In New Mexico, legislators received a salary of $8,460 in 2005, according to the Council of State Governments. This makes New York a “professional” legislature, unlike many other states that have a “citizen” legislature that pays minimal salaries and meets less often. In nearly every state, the Latino population has substantially grown in the past twenty years. Term limits were not enacted until the mid 1990s in most states. During the time associated with this study, the number of Latinos serving in legislatures has increased, although not to the extent the population has. That Latinos are political newcomers visà-vis other groups such as African-Americans is indisputable.

Effects of Population, Turnover, and Term Limits

41

Methods I test these hypotheses by analyzing data in presidential election years from 1992 to 2004. No states had passed legislative term limits prior to 1990. Data regarding the percentage of Latino citizens in each state is from the United States Census Bureau5. I computed the percentage of Latinos serving in each legislature by consulting the National Directory of Latino Elected Officials published annually. Moncrief, Niemi, and Powell generously provide data involving term limits and turnover based on their article published in August 2004. Data measuring the extent of professionalization come from Squire (2007).6 I obtained data on citizen ideology first developed by Erickson, Wright, and McIver (1993) and later refined by Wright (2004). Because of Nebraska’s unicameral legislature, I code it as an upper chamber. I conduct the analysis for the two chambers combined with a dummy variable for the chamber. I analyze each legislature beginning with the 1992 presidential election, every four years until 2004. Variables The dependent variable is the percentage of Latinos serving in each chamber in each state legislature. I have identified Latino legislators in each chamber by consulting the National Association of Latino Elected and Appointed Officials (NALEO).7 The United States Census Bureau collected the demographic data. The demographic variables are the percentage of a given state’s population who are identified in the most recent census as Latino citizens. 5

6

7

The percentage of Latino citizens in a state as collected by the Census Bureau is the most accurate measure available comparing the Latino population. The percentage of Latino voters in a state is not included in this analysis for a couple of reasons. First, I am interested in the connection between the Latino population of a state and the influence of Latinos in state legislatures. Second, just because some Latinos are not registered to vote, do not vote, or are not eligible to vote does not mean that they cannot, do not, and/or should not have influence in the political process. Members of Congress have been shown to be particularly sensitive to the potential preferences of people in their districts (Arnold 1990). Politicians are therefore quite aware that many Latinos will register and eventually vote, thus changing the demographics of districts all across the country. There is no significant correlation between professionalization and percentage Latino in a state. Whereas California has a professional legislature, Florida, New Mexico, and Texas have citizen legislatures. That is, highly professional legislatures are not always in medium to large states. I have confirmed NALEO’s listing of Latino legislators by comparing their list with the National Hispanic Caucus of State Legislators (NHCSL) roster.

42

Latino Representation in State Houses

The explanatory variables also include the percentage of Latino citizens in the population for all fifty states and the entire nation. I use U.S. Census data collected from the American Community Survey to estimate the percentage of Latino citizens in the district, as well as estimates from the William Velazquez Institute (see http://www.wcvi.org). A dummy variable is included for the upper chamber (1 if upper chamber; 0 otherwise). Two dummy variables are also included for states with term limits. The first term limit variable measures the year term limits were enacted. For example, if term limits were enacted in 1992 for a particular state, this observation is coded for that year and any year afterward until implementation. The rationale here is that the enactment of term limits may have an anticipatory effect on legislators, who may decide to alter their career plans based on future limits. The second term limit variable measures the year term limits took effect. This variable measures directly the effect of term limits on those legislators who are forced out of office by term limits. For example, if term limits became effective in a given state in 2000, then that observation is coded 1 for 2000 and 2004. For states that had term limits enacted or carried out in off years, such as 1995, then the observation is coded 1 for the subsequent year. States that initially passed term limits for legislators but were subsequently repealed either by the courts or by the legislature are coded 0. These states include Oregon, Idaho, and Utah. In addition, I track the percentage turnover in the lower and upper chambers for all fifty states and Congress every four years along presidential election years. I perform a feasible generalized least squares (FGLS) time-series regression in order to estimate the effects of the explanatory variables on the percentage of Latinos in legislatures over the time period from 1992 through 2004, every four years on the presidential election cycle. Because the repeated observations in the dataset are not independent, an FGLS model instead of an Ordinary Least Squares (OLS) model is necessary to avoid biased estimates. In particular, I use FGLS to address the problem of serial correlation evident in this dataset as it is in most time-series data (Worrall and Pratt 2004). That is, the percentage of Latinos serving in the legislature in 1992 is correlated with the percentage of Latinos serving in legislatures in 1996, 2000, and 2004. The Parks method is not used because of the many problems associated with the overconfidence of standard error estimates (Beck and Katz 1995).8 For the sake 8

The Parks method is feasible generalized least squares with the elimination of serial correlation of errors by examining the residuals to estimate “unit-specific serial correlation

Effects of Population, Turnover, and Term Limits

43

of robustness, these models are also estimated using random and fixed effects generalized least squares (FGLS), as well as OLS, and for the most part the substantive results remain the same (more on this later). The percentage Latino in the population is based on the 1990 Census for 1992 and 1996, whereas the updated Census 2000 numbers are used for 2000 and 2004. Institutional and Demographic Determinants of Latino Representation As shown in Table 2.2, [T]here is a strong relationship between the percentage of Latino citizens in a state and the percentage of Latinos in the legislature. A bivariate model provides strong support for the demographic hypothesis.9 It is still undoubtedly the case that Latino legislators are elected from states that have higher percentages of Latino citizens and even more so from districts that are majority-Latino. The overall percentage of Latinos in a state is also strongly related to the percentage of Latinos in the state legislature, although the relationship is not as strong.10 This chapter also explores the conditions under which Latinos are elected to seven of the nine state legislatures with the largest Latino memberships, and the next chapter will provide precise estimates of district-level effects. A multivariate regression including the two dummy variables for the enactment and implementation of term limits legislation is performed for 1992, 1996, 2000, and 2004.11 The percentage of Latino citizens in a state remains a significant predictor. There is no evidence to suggest that Latino representation is affected in any way by the enactment of term limits legislation. The percentage turnover has positive but insignificant effects on Latino representation. As the percentage of new legislators rises in a state, the percentage of Latinos in a legislature also rises. Some older legislators may have retired, but it is difficult to conclude, based on the term limit dummy variable, that term limits in general have had a positive effect on Latino representation. It is the case, however, that the opening up of otherwise entrenched seats has a positive effect on Latino

9 10

11

of the errors, which are then used to transform the model into one with serially independent errors.” See Beck and Katz (1995) for more on the different approaches to TSCS estimations. Appendix A shows the results of several other model specifications. This is based on a separate bivariate analysis not shown in Table 2.2 but available upon request. The N is 368, and not 396, because of some missing percentage turnover data in Moncrief et al’s dataset.

44

Latino Representation in State Houses

Table 2.2. The Effects of Turnover and Term Limits on Latino Representation, 1992, 1996, 2000, and 2004 Variables Percentage Latino Citizens in State Enactment of Term Limits Implementation of Term Limits Percentage Turnover Chamber State Ideology Professionalization Professionalization X Percentage Latino Citizens in State Term Limits Imp X Percentage Latino Citizens in State Size of Legislature Constant N Years Included

.965*** (.048) .071 (.367) -1.09* (.512) .017 (.009) -.115 (.280) .013** (.005) 1.59 (2.16) -.563** (.199) .348*** (.057) (1.78) .001 (.003) -2.45*** (.669) 368 1992 1996 2000 2004

Notes: Entries are regression coefficients; standard errors in parentheses. Feasible Generalized Least Squares Cross-Sectional Time-Series Regression. *p